KNOWBEL TECHNOLOGIES

Title

KNOWBEL TECHNOLOGIES : Get knowledge out of information !

Abstract

Access to information and the fruitful use of information are two central concerns for businesses. In its usual form, text information (press releases, announcements, meeting notes, Web pages, mail) is not structured (or has very little structure) and is therefore difficult to access when there are large volumes to be managed. In addition, under these conditions it is difficult to relate separate items of information to each other. The result is that the mine of knowledge constituted by the text databases produced or managed by businesses is currently underexploited.

KnowBel Technologies will specialise in the creation of software and the supply of high value-added services aiming to produce structured information (of ‘knowledge’) from non-structured text information.

From a more technical point of view, the methods employed derive from data retrieval and from knowledge engineering. Linguistic engineering tools are used to analyse texts, extract precise elements of information and place these in relation to one another. Among other things, this involves:

• assembling scattered and fragmented items of information
• matching up items of information that have been formulated in different ways
• relating items of information to one another so as to produce new knowledge

The outlook for economic exploitation of the research results

The first products and services to be marketed by the spin-off will be based on B-Ontology, a biographical knowledge base that is automatically updated and that has been developed over the last two years as part of an IRSIB project that has linked Cental with the Presse Belga press agency. These first products will be aimed mainly (but not exclusively) at the media sector (printed and audio-visual news media) as well as at ‘content suppliers’ on the Internet.

The spin-off company will then aim to broaden the scope of its activities and to offer made-to-measure data extraction services for businesses and public institutions. While we are not the only ones to have noted this need for applications dedicated to the intelligent handling of information, the originality and power of our offer (compared to that of our potential competitors) rests, to a great extent, on the promotion of hybrid solutions which, simultaneously offer a mix of linguistic relevance and statistical adaptability.